Part Select RAG Bot - AI Assistant Chrome Extension

Overview

The RAG Bot is an AI-powered Chrome Extension that assists users with inquiries about appliance parts from partselect through a dynamic, interactive chat interface. The project integrates web scraping, data processing, vector search databases, and an AI-driven query handling system to deliver accurate and context-aware responses.

Project Structure

Key Components and Their Roles

Web Scraper (Puppeteer.js)

• Path: part_select/Puppeteer.js
• Functionality: Automates the web browsing to scrape appliance parts data from online catalogs.
• Technologies Used: Puppeteer for controlling headless Chrome to scrape web data.
• Execution: Run periodically or manually to update the dataset.

Data Processing Script (data_processing.py)

• Path: parts-select/src/components/Models/data_processing.py
• Functionality: Transforms raw scraped data into structured JSON, generates embeddings using OpenAI's models, and uploads these embeddings to Pinecone for vector-based searching.
• Technologies Used: Python, OpenAI, Pinecone.
• Execution: Run after scraping to process new data and update the vector database.

Query Handling Module (query_handling.py)

• Path: parts-select/src/components/Models/query_handling.py
• Functionality: Handles user queries from the Chrome Extension, processes them to extract meaningful information, performs searches against the Pinecone database, and formats responses.
• Technologies Used: Python, Pinecone, Regex for direct part number extraction.
• Execution: Integrated within the server; processes queries in real-time as they are received from the frontend.

Server (server.py)

• Path: parts-select/src/components/Models/server.py
• Functionality: Acts as the backend server for the Chrome Extension, handling HTTP requests, interfacing with the query handling module, and sending responses back to the frontend.
• Technologies Used: Flask, Flask-CORS.
• Execution: Must be running whenever the Chrome Extension is in use to handle API requests.

Frontend UI (ChatWindow.js)

• Functionality: Provides the interactive user interface in the Chrome Extension where users can send queries and view responses.
• Technologies Used: React.js, CSS for styling.
• Execution: Loaded in the user’s Chrome browser as part of the extension.

Setup and Execution Instructions

Prerequisites

• Node.js (v14 or later)
• Python (v3.8 or later)
• Yarn or npm for managing JavaScript dependencies
• Access to OpenAI and Pinecone services, with API keys

Installation Steps

1. Clone the project repository:

   git clone https://github.com/your-username/parts_select_final
   
   

2. Setup the Backend
   cd /path/to/server/directory pip install -r requirements.txt

Configure .env file with necessary API keys

3. Prepare and Run the Data Processing Scripts python parts-select/src/components/Models/data_processing.py

4. Run the Server python parts-select/src/components/Models/server.py

5. Setup and Launch the Frontend cd /path/to/frontend/directory yarn install yarn build

Load the build folder as an unpacked extension in Chrome

6. Running the Scraper node part_select/Puppeteer.js